Astronomy is all about data. the universe grows, so does the information we gather about it. Some of the biggest challenges for the next generation of astronomy lie in how we will study all the data we collect. To address these challenges, astronomers are turning to machine learning (ML) and artificial intelligence (AI)to create new tools to quickly make the next big discoveries. In this article, we’ll look at four ways AI is helping astronomers study and understand the universe.
- Planet hunting: There are several ways a planet can be discovered. However, the most successful is the study of transits . When an exoplanet passes in front of its parent star, it blocks some of the light that we can see. By observing many orbits of an exoplanet, astronomers create a picture of the dips in the light, which they can use to determine the properties of the planet – such as its mass, size and distance from its star . NASA’s Kepler space telescope has used this technique with great success, monitoring thousands of stars at once and observing the dips caused by the planets.
Read also: NASA: We are looking for habitable planets and intelligent life forms
Humans are pretty good at spotting these dips, but it’s a skill that takes time to develop. With many missions dedicated to finding new exoplanets, like NASA’s Transiting Exoplanet Survey Satellite, humans can’t keep up. That’s where AI comes in. Time series analysis techniques — which analyze data as a sequential sequence over time — have been combined with a type of AI to successfully identify exoplanet signals with up to 96% accuracy.
- Gravitational waves: Time-series models are not only great for finding exoplanets, but also for finding the signals of the most cataclysmic events in the universe – mergers between black holes and neutron stars. When these dense bodies fall inward, they send ripples in spacetime that can be detected by measuring faint signals here on Earth. The LIGO and Virgo have detected the signals of dozens of such events with the help of machine learning. By training models on simulated black hole, the LIGO and Virgo teams can spot any events as they happen and send alerts to astronomers around the world to point their telescopes in the right direction.
- The changing sky: When the Vera Rubin Observatory , currently under construction in Chile , comes online , it will survey the entire sky every night—collecting more than 80 terabytes of images at a time —to see how stars and galaxies in the universe vary over time. A terabyte is 8,000,000,000,000 bits. During its scheduled operations, Rubin’s Legacy Survey of Space and Time (LSST) will collect and process hundreds of petabytes of data . 100 petabytes is about the amount of space it takes to store every photo on Facebook, or about 700 years of full-length high-definition video. You won’t be able to just log on to the servers and download that data, and even if you did, you wouldn’t be able to find what you’re looking for. Machine learning techniques will be used to search these next-generation surveys and highlight important data.
See also: Will NASA's Nancy Grace Roman Telescope find 100,000 planets?
- Gravitational lensing: As we collect more and more data about the universe, we sometimes have to edit or even throw out data that isn't useful. So how can we find the rarest objects in these data sets? One celestial phenomenon that excites many astronomers is strong gravitational lensing. This happens when two galaxies align along our line of sight and the gravity of the closer galaxy acts as a lens and magnifies the more distant object, creating rings, crosses, and double images. Finding these lenses is like finding a needle in a haystack. It's a quest that will become more difficult as we collect more and more images of galaxies.
In 2018, astronomers from around the world participated in the Strong Gravitational Lens Finding Challengeto see who could make the best algorithm for automatically finding these lenses.
Proposal: Universe: The most distant explosion ever known has been discovered
The winner of this challenge used a model called a convolutional neural network, which learns to analyze images using different filters until it can classify them based on whether they contain a lens or not. These models proved to be even better than humans, finding subtle differences in images that we humans have difficulty noticing.
Source of information: freepressjournal.in
